A basic tenet of all theories of reading comprehension is that the representation of a text in memory includes more than explicit text propositions. Inferences are assumed to be drawn and integrated into the representation. However, there are disagreements about which types of inferences occur. Experimental evidence relevant to this issue is mixed. Therefore, one series of experiments is proposed to provide stronger tests of inferential processes during reading, giving separate consideration to several classes of inferences. A second series of experiments will investigate the consequences of inferencing during reading upon memory for text. Such inferences require cognitive resources which might otherwise be committed to rehearsal, and to encoding of new inputs. On the other hand, inferencing involves prolonged attention to certain text propositions, review of information read earlier, and the integration of text propositions. These processes should aid text memory. We plan to examine text recall as a function of inferential procesing requirements during reading. Our long-range goal is a theory capable of specifying the inferences readers will draw while reading, the conditions influencing the activation and integration of such inferences, and the effects of these inferential activities upon memory of specific text propositions. The planned research is a beginning toward this goal.